REST API: sources and ingestion
Manage knowledge sources and observe asynchronous retraining work.
Source automation follows the same lifecycle as the dashboard: create or discover a source, wait for its processing state, then retrain when the relevant knowledge is ready. Website and page crawling are different operations, while files first obtain an upload URL and then request processing for the returned file identifier. Keep file bytes and API credentials on trusted infrastructure.
Polling is expected for asynchronous ingestion. Request retrain logs only as frequently as an operator needs to observe progress, and use bounded backoff rather than repeatedly restarting a crawl or retrain. A successful submission means work was accepted; it is not proof that every page, file, or external connection has finished indexing.
List and delete operations are source-type-specific. Do not substitute a website identifier for a page, Q&A, file, or Notion source identifier. Before a destructive cleanup, confirm which source the agent currently uses and capture enough metadata for a human to recover from an accidental removal. Consult the generated reference for accepted crawl and upload payloads.